MPI-INF Logo
Campus Event Calendar

Event Entry

What and Who

Image Classification with Limited Training Data and Class Ambiguity

Maksim Lapin
Max-Planck-Institut für Informatik - D2
Promotionskolloquium
AG 1, AG 2, AG 3, AG 4, AG 5, RG1, SWS, MMCI  
Public Audience
English

Date, Time and Location

Monday, 22 May 2017
14:00
60 Minutes
E1 4
024
Saarbrücken

Abstract

Modern image classification methods are based on supervised learning algorithms that require labeled training data. However, only a limited amount of annotated data may be available in certain applications due to scarcity of the data itself or high costs associated with human annotation. Introduction of additional information and structural constraints can help improve the performance of a learning algorithm. In this talk, we study the framework of learning using privileged information and demonstrate its relation to learning with instance weights. We also consider multitask feature learning and develop an efficient dual optimization scheme that is particularly well suited to problems with high dimensional image descriptors.


Scaling annotation to a large number of image categories leads to the problem of class ambiguity where clear distinction between the classes is no longer possible. Many real world images are naturally multilabel yet the existing annotation might only contain a single label. In this talk, we propose and analyze a number of loss functions that allow for a certain tolerance in top k predictions of a learner. Our results indicate consistent improvements over the standard loss functions that put more penalty on the first incorrect prediction compared to the proposed losses.

Contact

Connie Balzert
9325-2000
--email hidden
passcode not visible
logged in users only

Connie Balzert, 05/09/2017 11:12 -- Created document.